<p>As a type of pipeline robot, this paper designs an intelligent pipeline inspection gauge (PIG) that integrates an optimized sealing cup design with a strain sensing system and a virtual reality system for detecting dent defects in the pipeline. The mechanical properties of a ordinary cup and a high-passability cup when PIG runs through the dent are compared using the finite element method, and the effect of dent size on the strain distribution on the edge line of the high-passability cup is analyzed. The study showed that the strain peak value showed a trend of increasing with the increase in the dent depth. We developed an embedded strain sensing system within the cup based on STM32 microcontroller, enabling the cup to detect pipeline dent defects while maintaining high-passability. Experimental validation using a dent defect with 13% of the outer diameter (13% OD) confirmed the consistency between and the simulation data and the collected strain signals. A virtual reality system was developed in Unity 3D engine. The collected strain data can be displayed in real time and visualized in Unity Graphical User Interface (UGUI), and the strain signal is mapped to the position of the pipeline robot. A multilayer perceptron (MLP) neural network prediction model is built to accurately predict the strain signals of unknown defects. The proposed model achieved a prediction accuracy of 95.6% on the prediction set and demonstrated superior generalization capability compared to conventional regression approaches, providing reliable support for pipeline risk assessment and defect quantification.</p>